Stock market prediction using Long Short Term Memory algorithm combined with Genetic Algorithm
Keywords:
Genetic Algorithm, LSTM, Prediction, Stock, Nifty.Abstract
In this article, Long Short Term Memory (LSTM) algorithm with optimized value of windows size and number of units was used to predict the Nifty – IT daily stock values. Genetic algorithm technique was utilized for optimization. Window Size=36, Number of Units=2 was found to be the best suitable values for prediction using LSTM. Various parameters such as cosine similarity (CS), Root mean square Error (RMSE), R2 etc. were used to determine the quality of the prediction done. The value of CS, RMSE and R2 was calculated as 0.99992, 0.00374 and 0.97713 respectively. The CS and R2 values are closed to 1.0, whereas the value of RMSE is very small, which is an indicative that the actual and predicted data are much closed to each other.
References
- https://scripbox.com/pf/what-is-nifty/ Accessed on (01.08.2022)
- https://www.niftyindices.com/indices/equity/sectoral-indices/nifty-it (Accessed on 01.08.2022)
- D Kumar, P K Sarangi, and R Verma, Materials Today: Proceedings, https://doi.org/10.1016/j.matpr.2020.11.399 (2020)
- Kim Y, Roh JH, Kim H, Sustainability 10:34–53 (2017)
- Armano G, Marchesi M, Murru A, Inf Sci 170:3–33 (2005)
- Pal SK, Wang PP, Genetic algorithms for pattern recognition. CRC Press, Boca Raton, p 336 (1996)
- Santra SA, Lin LJ Integrating long short-term memory and genetic algorithm for short-term load forecasting. Energies 12:2040 (2019). https://doi.org/10.3390/en12112040
- Thakur N, Karmakar S, and Soni S, Int. j. inf. tecnol.,14(4):1961–1966 (2022)
- https://www.kaggle.com/datasets/sudalairajkumar/nifty-indices-dataset (Accessed on 01.08.2022)
Downloads
Published
Issue
Section
License
Copyright (c) IJSRCSEIT
This work is licensed under a Creative Commons Attribution 4.0 International License.